From 1d4c212441d5162c05b4968ff9c64ee0219bf44b Mon Sep 17 00:00:00 2001
From: Oneric <oneric@oneric.stub>
Date: Wed, 15 May 2024 01:20:27 +0200
Subject: [PATCH] dbprune: shortcut array activity search

This brought down query costs from 7,953,740.90 to 47,600.97
---
 lib/mix/tasks/pleroma/database.ex        | 16 +++++++++++++++-
 test/mix/tasks/pleroma/database_test.exs |  4 ++++
 2 files changed, 19 insertions(+), 1 deletion(-)

diff --git a/lib/mix/tasks/pleroma/database.ex b/lib/mix/tasks/pleroma/database.ex
index b8f19551a..ac0f15a33 100644
--- a/lib/mix/tasks/pleroma/database.ex
+++ b/lib/mix/tasks/pleroma/database.ex
@@ -36,6 +36,19 @@ defmodule Mix.Tasks.Pleroma.Database do
         ""
       end
 
+    # Activities can either refer to a single object id, and array of object ids
+    # or contain an inlined object (at least after going through our normalisation)
+    #
+    # Flag is the only type we support with an array (and always has arrays).
+    # Update the only one with inlined objects.
+    #
+    # We already regularly purge old Delete, Undo, Update and Remove and if
+    # rejected Follow requests anyway; no need to explicitly deal with those here.
+    #
+    # Since there’s an index on types and there are typically only few Flag
+    # activites, it’s _much_ faster to utilise the index. To avoid accidentally
+    # deleting useful activities should more types be added, keep typeof for singles.
+
     # Prune activities who link to a single object
     {:ok, %{:num_rows => del_single}} =
       """
@@ -61,7 +74,8 @@ defmodule Mix.Tasks.Pleroma.Database do
       delete from public.activities
       where id in (
         select a.id from public.activities a
-        join json_array_elements_text((a."data" -> 'object')::json) as j on jsonb_typeof(a."data" -> 'object') = 'array'
+        join json_array_elements_text((a."data" -> 'object')::json) as j
+             on a.data->>'type' = 'Flag'
         left join public.objects o on j.value = o.data ->> 'id'
         left join public.activities a2 on j.value = a2.data ->> 'id'
         left join public.users u  on j.value = u.ap_id
diff --git a/test/mix/tasks/pleroma/database_test.exs b/test/mix/tasks/pleroma/database_test.exs
index 18b4752bf..e92ffbf98 100644
--- a/test/mix/tasks/pleroma/database_test.exs
+++ b/test/mix/tasks/pleroma/database_test.exs
@@ -478,6 +478,7 @@ defmodule Mix.Tasks.Pleroma.DatabaseTest do
       |> Map.merge(%{
         local: false,
         data: %{
+          "type" => "Flag",
           "id" => "remote_activity_existing_object",
           "object" => ["non_ existing_object", "existing_object"]
         }
@@ -488,6 +489,7 @@ defmodule Mix.Tasks.Pleroma.DatabaseTest do
       |> Map.merge(%{
         local: false,
         data: %{
+          "type" => "Flag",
           "id" => "remote_activity_existing_actor",
           "object" => ["non_ existing_object", "existing_actor"]
         }
@@ -498,6 +500,7 @@ defmodule Mix.Tasks.Pleroma.DatabaseTest do
       |> Map.merge(%{
         local: false,
         data: %{
+          "type" => "Flag",
           "id" => "remote_activity_existing_activity",
           "object" => ["non_ existing_object", "remote_activity_existing_actor"]
         }
@@ -508,6 +511,7 @@ defmodule Mix.Tasks.Pleroma.DatabaseTest do
       |> Map.merge(%{
         local: false,
         data: %{
+          "type" => "Flag",
           "id" => "remote_activity_without_existing_referenced_object",
           "object" => ["owo", "whats_this"]
         }